跳到正文
arXiv cs.AI· Xintong Wang, Haihan Nan, Ruidong Li·· 4 小时前AI 评分33

多模态时间序列预测诊断基准:文本与历史一致、冲突或未指明的首次评测

Aligned, Conflicting, or Unspecified? A Diagnostic Benchmark for Multimodal Time Series Forecasting

AI 导读

一篇被 IEEE AIoT 2026 Workshop 录用的 arXiv 论文提出首个按文本-历史关系组织的真实世界多模态时间序列预测诊断基准,包含 1,762 个文本-窗口对,覆盖 Weather、Wikipedia 与 Finance 数据流,标签分为 Aligned、Conflicting 或 Unspecified。

正文 · 原文

View PDF HTML (experimental)

Abstract:Artificial Intelligence of Things (AIoT) systems increasingly combine sensed time series with external text such as expert reports and news, yet how reliably models use text that agrees or conflicts with the sensed history remains unclear. We introduce the first real-world diagnostic benchmark organized by text-history relationship: 1,762 text-window pairs across sensor telemetry (Weather), web-service traffic (Wikipedia), and market data streams (Finance), each labeled Aligned, Conflicting, or Unspecified by whether the text's direction agrees with the historical trend, opposes it, or is not stated. Across ten pretrained models evaluated zero-shot, we investigate text selection and combination under text-history relationships (RQ1) and error diagnosis and horizon-scaled revision control (RQ2). For RQ1, we compare two types of text on paired Weather windows: one directly forecasts temperature changes, while the other describes related weather conditions, such as winds and fronts. Under Aligned, the first type yields the lowest error for most models. Under Conflicting, it increases most models' error above the no-text baseline; we therefore add only its stated magnitude and timing to the second type. Compared with either type alone, this combination lowers mean absolute scaled error (MASE) for six models and raises directional accuracy (DA) for seven. For RQ2, an error decomposition attributes the error increase to oversized revisions, and a training-free rule that limits them outperforms forecasts both with and without text for the four models with this trade-off, lowering MASE by 4.49%-21.98% and raising DA by 2.99%-9.71%. Our dataset is available at this https URL.
Comments: Accepted at IEEE AIoT 2026 Workshop
Subjects: Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.17680 [cs.AI]
  (or arXiv:2611.17680v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.17680

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Xintong Wang [view email]
[v1] Fri, 13 Nov 2026 02:47:28 UTC (633 KB)

来源:arXiv cs.AI · arxiv.org